INDEX

Explanations

phrases relating to social or political issues

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

conom

-0.90

alysis

-0.89

imester

-0.87

ELY

-0.85

cosystem

-0.83

mediate

-0.82

tests

-0.82

etimes

-0.82

Reviewer

-0.80

 Flavoring

-0.79

POSITIVE LOGITS

 swast

1.47

 logo

1.36

 likeness

1.31

 adorned

1.31

 silhouette

1.29

 logos

1.28

 silhou

1.23

 painted

1.17

 inscribed

1.16

 insign

1.16

Activations Density 2.846%

phrases relating to social or political issues

No Comments

No Known Activations